Customer Story:

OpenWeb automates real-time content moderation with Qwak

OpenWeb improves the quality of conversations online by building an internet where content creators of every kind are empowered to truly own their audience relationships and thrive independently.
Qwak was brought aboard to support a growing team of data scientists with a goal to scale model quality and bridge the engineering gap often experienced with new data science groups.

<50ms
Real-Time
inference
Central
Model  
repository
5
New Models introduced to OpenWeb's stack
About Openweb
OpenWeb is building the healthy, social layer of the internet. OpenWeb partners with publishers and brands to build audience relationships.
Industry
Internet
Social
Publishing
Machine Learning
Use Case:
Moderation of content - Automated NLP models to moderate discussion quality.

Model Frameworks:
Pytorch
Qwak stack:
Build system
Serving/Hosting
Analytics
" People ask me how I managed to deploy so many models while onboarding a new team within a year.
My answer is: Qwak"
Idan Benaun
Director of Machine learning and Data Science @ OpenWeb
The Challenge
Data Science team expansion and need to show immediate value.
  • OpenWeb's growing data science team needed to show immediate value. Other solutions could not provide engineering support the team needed to move quickly and not have to deal with maintenance and configuration of servers and infrastructure.
  • Change management and basic configurations were very complex and difficult to execute on other solutions tested.
  • OpenWeb's team required real-time inference response going through multiple models at a speed of less than 50 ms.
  • Real time model deployment required.
How Qwak Assisted?
  • Qwak Build - Integration to existing models is very simple and intuitive, just adding a line of code using Qwak’s Python SDK. Data scientists with almost no engineering expertise can run integrations.
  • Qwak's Inference - NLP based modifications on inference are easy to execute in comparison to other solutions such as Sagemaker. Qwak allowed us to get results in real-time while keeping inference in low latency.
  • Qwak's Monitoring - Out of the box infrastructure monitoring, alerting and logs,   notifies OpenWeb's team on any issue and allows immediate troubleshooting.
  • Qwak's Analytics  allows OpenWeb to query model data and keep track of  prediction data.
  • Qwak's Support  was very helpful and gave us immediate assistance during the onboarding and moving to production process

OpenWeb ML Model Infrastructure mockup
"Qwak's service is impeccable and the platform provides  immediate feedback which allows us to turn things around quickly when needed. "
Idan Benaun
Director of Machine learning and Data Science @ OpenWeb

Experience Seamless MLOps Like Never Before

 
Try for Free